In VLSI semiconductor film processing, dielectric materials are used as insulating layers between various circuits and layers of circuits in integrated circuit devices and other related electronic devices. As technology advances, these devices become increasingly smaller and denser, which results in narrower spacing between adjacent circuit elements, such as conductors which carry rapidly switching transient signals. Consequently, films with lower dielectric constants are needed to avoid cross talk and capacitive coupling between conductors. These polymer-type films are used because they can effectively fill gaps between adjacent circuit elements.
A recent development in VLSI semiconductor film processing utilizes insulating polymer films with dielectric constants of approximately 2.5 and less. Such films are typically formed using CVD chamber systems comprising of a vaporization chamber, a pyrolysis chamber, and a deposition chamber. A dimer is first vaporized and then pyrolized, i.e., cleaved into a monomer vapor form. The monomer then enters the deposition chamber and condenses on the surface of a semiconductor wafer to form a linked polymer film on the wafer.
However, during the deposition process, the polymer film can also form undesirably on the edge and backside of the wafer. Problems arise because the polymer film does not adhere readily to the wafer edge. Thus, a tendency exists for the polymer film formed on the edge and backside of the wafer to chip and flake, thereby introducing contaminants into the deposition chamber during fabrication of the wafer. Wafer cassettes can also be contaminated during transfer of wafers from one processing step to another. The cassettes typically only contact the wafers at their edges, so that any polymer film on the edges may be scraped off during wafer transportation. Furthermore, because some applications require the wafer backside to be free from any film formation, both backside and edge exclusion are desired during film deposition.
Conventional techniques for excluding deposition on the backside of a wafer include etching the backside of the wafer after film deposition. This approach entails added expense and time to the fabrication process. Other techniques for excluding deposition on the edge of a wafer, such as those disclosed in U.S. Pat. No. 5,620,525 to van de Ven, et al. and U.S. Pat. No. 5,556,476 to Lei et al., inject an inert control gas along the wafer edge from underneath the wafer up into the deposition chamber during tungsten CVD to prevent edge deposition. However, these methods are not effective for use in the monomer deposition process. One reason is the differences between monomer deposition and tungsten deposition, i.e., the flow rate of the monomer and the pressure of the deposition chamber are both low, typically in the range of 2-7 sccm and on the order of 50 mTorr, respectively, and the molecular weight of the monomer molecules is very high, i.e. 176 a.m.u. Because the flow rate of process gas for tungsten deposition is high, i.e., 2-3 standard liters per minute, and the molecular weight of the process gas is low, a high flow rate, i.e., 300 sccm to 3 standard liters per minute, of inert gas across the wafer edge can effectively prevent the process gas from depositing on the wafer edge. However, when the process gas flow is low, i.e., for monomer deposition, a high flow rate of the inert gas into the deposition chamber can dilute the process gas, resulting in substandard film deposition. By decreasing the inert gas flow rate, the inert gas is not able to adequately prevent high molecular weight process gases from depositing on the wafer edge.
Accordingly, a process and structure are desired which excludes polymer film formation on the edge and backside of a wafer during the deposition process.